This paper investigates the affordances of ChatGPT in higher education and examines how artificial intelligence (AI) technologies may reshape the learning function in higher education. This study utilizes a grounded theory approach to analyze data collected from six sessions of panel discussions with participants from US universities. The study presents findings in the form of a framework encompassing four categories of affordances that "mitigate challenges in traditional learning environments," "enhance effective educational practices," "transform traditional learning approaches," and "negatively impact current effective educational practices." The framework, including the affordances and sub-affordances, illustrates the potential impact of this technology on the learning function of higher education. This research contributes to the literature by establishing a foundation for understanding ChatGPT's role in higher education and fostering further inquiry. Additionally, the findings can assist higher education decisionmakers in formulating policies and strategies to capitalize on the opportunities presented by AI tools like ChatGPT while mitigating potential risks.